Performative Prediction in Time Series: A Case Study

Abstract

Performative prediction is a phenomenon where a model's predictions, or the decisions based on these predictions, may influence the outcomes of the model. This is especially conspicuous in a time series prediction setting where interventions occur before outcomes are observed. These interventions dictate which data points in the time series can be used as inputs for future predictions. In this paper, we represent patient-reported symptom values collected during their oncology appointments as a time series. We use a decision-tree based model to predict a patient's future symptom values. Based on these predictions, clinicians decide which symptom values will be observed in the future. We propose methods to provide robustness against the problem of performative prediction in time series. Our results characterise how performative prediction may lead to a 29.4% to 40.7% higher error across different symptoms.